Directory UMM :Data Elmu:jurnal:I:Insect Biochemistry and Molecular Biology:Vol31.Issue2.Feb2001:

(1)

www.elsevier.com/locate/ibmb

Immune response of

Drosophila melanogaster

to infection with the

flagellate parasite

Crithidia

spp.

Nathalie Boulanger

a,*

_{, Laurence Ehret-Sabatier}

_{, Reto Brun}

_{, Daniel Zachary}

_,

Philippe Bulet

_{, Jean-Luc Imler}

a_{Re´ponse Immunitaire et De´veloppement chez les Insectes, UPR 9022 du CNRS, Institut de Biologie Mole´culaire et Cellulaire, 15 rue Rene´} Descartes, 67000 Strasbourg, France

b_{Swiss Tropical Institute, PO Box, 4002 Basel, Switzerland}

Received 21 January 2000; received in revised form 20 March 2000; accepted 26 April 2000

Abstract

Insects are able to recognize invading microorganisms and to mount an immune response to bacterial and fungal infections. Recently, the fruitfly Drosophila melanogaster has emerged as a promising invertebrate model to investigate innate immunity because of its well-characterized genetics. Insects are also vectors of numerous parasites which can trigger an immune response. We have investigated the interaction ofDrosophila melanogasterwith the flagellate protozoanCrithidiaspp. We show that a per os parasitic infection triggers the synthesis of several antimicrobial peptides. By reverse phase HPLC and mass spectrometry, peptides were shown to be present in the hemolymph and not in the gut tissue, suggesting the presence of immune messengers between the site of the infection, namely the gut, and the fat body, the main site of synthesis for antimicrobial peptides. Interestingly, we have identified one molecule which is specifically induced in the hemolymph after infection withCrithidia, but not with bacteria, suggesting thatDrosophilacan discriminate between pathogens. When flagellates were injected into the hemolymph, a low synthesis of antimicrobial peptides was observed together with phagocytosis of parasites by circulating hemocytes. The data presented here suggest that Drosophila–Crithidia spp. represents an interesting model to study host defense against protozoan parasites.2001 Elsevier Science Ltd. All rights reserved.

Keywords:Insect immunity;Drosophila melanogaster;Crithidia; Flagellate parasite; Antimicrobial peptides

1. Introduction

Insects proliferate in diverse biotopes and are con-stantly exposed to various microorganisms. Conse-quently, during evolution they have developed efficient immune defenses to resist hostile environments. These defenses rely on different mechanisms: phagocytosis, activation of proteolytic cascades such as coagulation and melanization, and production of various antimicro-bial peptides (for reviews, see Richman and Kafatos, 1995; Hoffmann and Reichhart, 1997; Boman, 1998).

The Drosophila host defense has been particularly well

investigated and has served to elucidate the phylogenetic

* Corresponding author. Tel.: +333-88-417095; fax: +333-88-606922.

E-mail address:N.Boulanger@ibmc.u-strasbg.fr (N. Boulanger).

basis of innate immunity (reviewed in Hoffmann et al., 1999). Indeed, in response to an experimental systemic infection with bacteria,Drosophila synthesizes in its fat body (a functional equivalent of the vertebrate liver) a large panel of antimicrobial peptides. Despite the many different structures, all the gene-encoded antimicrobial peptides studied to date are small cationic molecules. Their size is mostly below 10 kDa and interestingly, they possess a broad spectrum of potent antimicrobial activity. Drosomycin has exclusively antifungal proper-ties, in contrast to drosocin, diptericin, cecropin and defensin which affect bacterial growth (Bulet, 1999). Several insect species are vectors of medically important parasitic diseases. Studies in mosquitoes have demon-strated that the presence of filarial worms or malaria parasites can induce an immune response in Aedes spp. (Beerntsen et al., 1994) and in Anopheles spp. (Dimopoulos et al., 1998), respectively. We were

(2)

inter-ested in developing a Drosophila model of parasitic infection to benefit from the genetic tools of this organ-ism. In particular, we wanted to compare the effects of parasitic infection to the well-studied antibacterial and antifungal responses. Drosophila has been reported to harbor kinetoplastida flagellates which include several species of the genus Crithidia (Wallace, 1966; Ismaeel, 1994). Crithidia parasites exclusively infect invert-ebrates and predominantly insects (Wallace, 1979). In general, flagellates develop in the digestive tract and interact with the intestinal epithelium by their flagellum. This interaction induces the formation of hemidesmo-somes (Molyneux and Killick-Kendrick, 1987) and fre-quently leads to damage of the intestinal cells (Schaub, 1994). Crithidia spp. parasites vary widely in their degree of host specificity (Wallace, 1966). As the species infecting Drosophila have not been identified, we have decided to use in this studyC. bombiandC. fasciculata, isolated from bumblebee and mosquitoes, respectively. Here we report thatper osinfection ofCrithidiatriggers a defense reaction in Drosophila. In particular we have observed, by several biochemical techniques, the induc-tion of several known antimicrobial peptides as well as the induction of new Drosophila Immune induced M ol-ecules (DIMs — Bulet and Uttenweiler-Joseph, 1999) in the hemolymph of gut-infectedDrosophilaadults. As

Crithidia parasites do not cross the gut barrier, our

results suggest that this tissue can signal to the fat body, the principal site of antimicrobial peptide synthesis. In contrast to oral infection which did not affect fly sur-vival, systemic infection had different effects on the flies depending on the Crithidiastrain used: whereas C.

fas-ciculata rapidly killed the flies,C. bombiwas harmless.

The role of phagocytosis by hemocytes and antimicro-bial peptides in the hemolymph was investigated to try to explain this difference.

2. Materials and methods

2.1. Drosophila strains

Oregon flies were maintained in the laboratory and used as a standard wild type strain of Drosophila mel-anogaster. The ILL 97 strain was collected from the field on fruit traps in the area of Strasbourg (France). The colony was amplified and maintained at 25°C on stan-dard corn meal medium. ILL 97 flies were crossed with Oregon to demonstrate that they belong to the

mel-anogaster species.

2.2. Parasite cultures

The two Crithidia species, C. fasciculata and C.

bombi were received from the Swiss Tropical Institute

(Basel, Switzerland).C. bombiwas isolated by Dr

Shyk-off (Zoological Institute, Basel, Switzerland) from the rectal ampula of local bumblebees,Bombus spp.C.

fas-ciculata was isolated from local mosquitoes,Culexspp.

by Dr Stemmberger (Vienna, Austria). Parasites were maintained in Schneider’s medium supplemented with 10% Fetal Calf Serum (FCS) containing 60 mg/l penicil-lin and 100 mg/l streptomycin at 25°C. Dense parasite populations were subcultured twice weekly.

2.3. Parasite infections

For per os infection, small cotton balls were soaked

with sterile parasite cultures. Adult flies from the two strains were fed on this medium for 24 h at room tem-perature. Control flies were maintained on sterile Schnei-der’s medium. Positive controls were provided by soak-ing a cotton ball with a mixture of bacteria,Escherichia coli(strain 1106) andMicrococcus luteus(strain A270). For systemic infection, parasites were washed with phos-phate buffer saline (PBS) and pelleted by centrifugation at 3000 rpm for 10 min. Parasite suspensions (4.6 nl — around 5000 parasites) were injected into the thorax of adult flies using a nanoinjector (Nanoject, Drummond Scientific Co., USA). To quantify the physical effect of the injury, a PBS injection was performed under the same conditions as for control flies.

2.4. Northern blot analysis

Total RNAs from whole flies were extracted and Northern blot experiments performed according to the procedures of Lemaitre et al. (1996). The following cDNA probes were used: diptericin, defensin, cecropin A1, drosomycin, drosocin and ribosomal protein 49 (rp49) as standard (Lemaitre et al., 1996). For each point, RNA was extracted from 20 flies. The intensity of the immune response was quantified by autoradiography for

per osinfection and with a Bio-Imaging Analyzer (BAS

2000, FUJIX, Japan) for parasitic injection.

2.5. Analysis of Drosophila hemolymph and extracts

from digestive tract by RP-HPLC and MALDI-TOF MS

The procedure used was derived from Uttenweiler-Joseph et al. (1998). Briefly, hemolymph of 30

Droso-philaadults infectedper oswith either (1) medium alone

(negative control), (2) a bacteria mixture (E. coliandM. luteus) or (3) parasites, was collected with the help of a nanoinjector and directly transferred in 50µl of acidified water (0.1% TFA) to reduce proteolysis. The sample was centrifuged for 10 min at 2000gto pellet hemocytes. Gut

tissue of 30 Drosophila adults were dissected, washed in PBS and transferred to acidified water. To lyse cells, samples were ground, sonicated and centrifuged. The supernatants were subjected to RP-HPLC on an

(3)

Aqua-pore RP 300 C8 column (1×100 mm, Brownlee). Hemolymph and gut extracts were separately analyzed by a linear gradient of 2–80% acetonitrile in 0.05% TFA over 80 min at a flow rate of 80 µl/min. The column effluent was monitored by absorbance at 214 nm and fractions were hand-collected. All HPLC purifications were performed with a Waters HPLC system equipped with a pump model 626, a controller model 600S and a detector model 486. For the subsequent analysis by MALDI-TOF-MS, fractions were selected according to the retention time of known antimicrobial peptides (Uttenweiler-Joseph et al., 1998) or according to major absorbance variations.

2.6. Electron microscopy

A parasite suspension of C. bombi or C. fasciculata

(4.6 nl corresponding to approximately 5000 parasites) was injected intoDrosophilalarvae of stage 3. At differ-ent time points following the injection (1 and 3 h), larvae were pricked to collect hemolymph and the hemolymph was centrifuged to pellet hemocytes and parasites. The pellet was fixed in a 4% solution of glutaraldehyde in 0.1 M phosphate buffer, pH 7.4. Preparations were post-fixed in 1% osmic acid for 1 h, then dehydrated and embedded in epoxy resin. Ultrathin sections were con-trasted with uranyl acetate–lead citrate and processed for transmission electron microscopy.

3. Results

3.1. Per os infection with flagellates induces expression of antimicrobial peptide genes in a field-strain of Drosophila but not in Oregon flies

Adult Oregon flies were orally infected with bacteria and parasites. Twenty four hours later, infection was confirmed by the presence of parasites, first detected in the diverticulum, then in the gut (data not shown). Infec-tion persisted for approximately 4 days after which the parasites were eliminated with the feces. To determine whether aper osinfection induced an immune response, we extracted RNA from whole flies 24 h post infection and probed Northern blots with cDNAs coding for

Dro-sophilaantimicrobial peptides. As shown in Fig. 1A, the

bacteria which had been orally administered induced a strong expression of drosocin in flies, whereas the two

Crithidia species did not induce a detectable expression

of drosocin messenger RNA in the Oregon strain. Simi-lar results were obtained with cecropin, diptericin, defen-sin and drosomycin probes (data not shown). As the Ore-gon flies had been maintained in laboratory culture conditions for several years, we suspected that they might have lost the capacity to react to the presence of parasites in the digestive tract. We therefore collected

Fig. 1. Antimicrobial peptide gene expression in flies following experimental per os infection. Flies from the Oregon (A) or ILL 97 (B) strains were orally infected with a mixture of Gram-positive and Gram-negative bacteria (bact.) or with twoCrithidiaspecies,C. bombi

(C.b.) or C. fasciculata (C.f.). Total RNA was prepared 24 h post-infection from control (c) or challenged adult flies, and analyzed by Northern blot for the induction of defensin (Def), cecropin A1 (CecA), diptericin (Dipt), drosocin (Drc) and drosomycin (Drom). The riboso-mal protein 49 (rp49) was used as quantitative control. A representa-tive experiment is shown.

flies in the field, and established a new line, ILL 97, that we exposed to per os parasite infection. As shown in Fig. 1B, theper os administration of bacteria induced a strong expression of the genes encoding cecropin, droso-cin, diptericin and drosomycin and to a lesser extent defensin. Significantly, these peptides were also induced by oral administration of Crithidiaparasites in this field strain, although the level of induction was lower than with bacteria. Induction by parasite infection was strong-est for drosocin and diptericin, whereas defensin was only marginally induced. We have subsequently used the ILL 97 flies to study the response ofDrosophilato Cri-thidiaspp. infection.

(4)

3.2. Per osinfection with flagellates triggers a systemic immune response

Surface epithelia have been shown to synthesize anti-microbial peptides both in mammals (reviewed in Ganz and Lehrer, 1998) and in various insect species, namely

Bombyx mori (Brey et al., 1993), Manduca sexta

(Russell and Dunn, 1996), Stomoxys calcitrans(Lehane et al., 1997), Anopheles gambiae (Dimopoulos et al., 1998) and D. melanogaster (Ferrandon et al., 1998; Ohresser and Imler, unpublished data). To analyze if

Crithidia induces antimicrobial peptide synthesis in the

digestive tract (local response) or in the fat body (systemic response), we performed a separate analysis by reverse phase HPLC (RP-HPLC) and mass spec-trometry of acidic extracts of gut tissue and hemolymph from individual flies afterper osinfection withCrithidia

spp. or with bacteria. As shown in Fig. 2, infection by

both C. fasciculata and bacteria induced a systemic

immune response illustrated by the appearance of droso-cin and drosomydroso-cin (peaks 1 and 2, respectively) in the hemolymph of infected flies, 24 h after infection. The

Fig. 2. Differential study ofDrosophila(ILL 97 strain) hemolymph after systemic orper osinfection. (A) Hemolymph from ILL 97 flies was collected 24 h post infection and analyzed by RP-HPLC. Flies were either injected with a mixture of Gram-positive and Gram-negative bacteria (inj. bact.), or infectedper oswith Schneider medium (control), a mixture of Gram-positive and Gram-negative bacteria (bact.) orC. fasciculata

(C.f.). The numbers (1,2) indicate peaks containing induced antimicrobial peptides, and the letters (A,B,C) indicate peaks containing induced molecules. (B) The fractions corresponding to peaks 1 and 2 were subjected to MALDI-TOF mass spectrometry to demonstrate that they contain drosocin and drosomycin, respectively. The results of the MALDI-TOF mass spectrometry analysis of peaks A, B and C are presented in Table 1. absence of defensin and cecropin in Fig. 2 may be explained by their early synthesis which reaches a maximum 7 h after infection (Bulet, unpublished data). We have made a similar analysis with the Oregon flies: drosomycin and drosocin were detected at considerably lower levels which confirmed the poor detection obtained for the Northern blot experiment (see above, Fig. 1A).

The per os infection with Crithidia spp. or bacteria

also induced molecules distinct from the known anti-microbial peptides (Fig. 2A). Mass spectrometry analy-sis of peak B yielded several masses ranging from approximately 3 kDa to 7 kDa, one of which (mol. wt 3175 Da) was specific for Crithidia infection (Table 1). The corresponding molecules are still under investi-gation. Interestingly, peak B was observed only in per os infected flies, and was not detected in flies injected with bacteria. Peaks A and C of Fig. 2A were only induced by bacteria, both afterper osadministration and injection. These data demonstrate that not only the nat-ure of infecting microorganisms but also the route of infection induce different molecules in the hemolymph

(5)

Table 1

Specificity of the immune response inDrosophilainjected into the thorax with bacteria or infectedper oswith different pathogens (bacteria and

Crithidiaspp.)a

Per os C. fasciculata

Peak Bacteria injection Per osbacteria infection Per os C. bombiinfection

infection

A DIM 1b _{DIM 1}

2 2

DIM 4 DIM 4

B 2 2899.0 2899.4 2899.0

3176.6 3175.2

4347.3 4347.8 4347.3

7193.0 7193.5 7193.0

C DIM 13 DIM 13 2 2

a_{The RP-HPLC fractions corresponding to induced peaks A, B and C were analyzed by MALDI-TOF-MS. DIMs 1 (1666 Da), 4 (1722}

Da), 13 (2651 Da) were previously described (Uttenweiler-Joseph et al., 1998). Measured molecular masses (MH+) are indicated for unknown induced molecules.

b _{DIM —}_Drosophila_{Immune-induced Molecules.}

of Drosophila. Using a similar methodology, we have

analyzed gut tissue from bacteria- andCrithidia-infected field strain Drosophila. We were unable to detect the presence of antimicrobial peptides in this type of extract 24 h after aper osinfection. These data were confirmed by reverse transcription-PCR (RT-PCR) analysis which showed expression of drosocin and diptericin in dis-sected fat bodies from per os infected flies, but not in dissected guts (data not shown).

3.3. Injection of C. fasciculata kills flies

Some flagellate parasites such as certainTrypanosoma

species naturally undergo part of their lifecycle in the insect hemolymph (Kaaya et al., 1986; Molyneux and Killick-Kendrick, 1987). Others invade accidentally hemolymph like Herpetomonas spp. or Blastocrithidia

bombi was not harmful and up to 80% of injected flies

had survived 6 days after injection of the parasite or PBS. Interestingly, dissection ofC. bombi-infected flies showed few viable (e.g. motile) parasites 4–6 days after injection while C. fasciculata-infected flies contained large numbers of highly motile parasites.

3.4. Injection of Crithidiaparasites induces a weak

systemic antimicrobial response

We next examined whether the antimicrobial peptides identified in theDrosophilahost defense against bacteria or fungi could participate in the humoral immune response after flagellate injection. Indeed, several authors in different parasite–insect systems have

Fig. 3. Survival curves ofDrosophilaILL 97 challenged with Crithi-diaspp. Groups of 15 flies were injected into the thorax with 4.6 nl of a suspension of around 5000C. bombi(C.b) orC. fasciculata(C.f) parasites and mortality was assessed at daily intervals. PBS was used as a negative control to measure the direct effect of the injection pro-cedure. Controls (C) were uninjected flies. Data represent means±SD of triplicates. A representative experiment is shown.

reported that antimicrobial peptides, when induced by a septic injury (Lowenberger et al., 1996) or when injected into the thorax (Shahabuddin et al., 1998), can control the establishment of parasite infection present in the hemolymph. It might be speculated that the difference in susceptibility of Drosophila to infection with either

C. fasciculata or C. bombi results from a difference in

the level of induction of antimicrobial peptides. To investigate this, we analyzed the expression of the corre-sponding genes in flies 1 and 2 days after injection of both parasites, and compared this response with that

(6)

induced by injection with bacteria (Lemaitre et al., 1996). As shown in Fig. 4, both parasites induced a simi-lar low level of expression of genes encoding diptericin, drosomycin and drosocin genes 24 h post injection. The levels were markedly lower than those induced by injec-tions of bacteria, and had returned to basal level or lower 48 h post infection.

3.5. Phagocytosis of Crithidiaparasites by hemocytes

In addition to antimicrobial peptides, circulating and sessile hemocytes contribute to the encapsulation of microorganisms during insect immune response (reviewed in Lackie, 1988). We therefore investigated whether C. bombi parasites, which disappear rapidly after injection into the thorax, were better internalized by phagocytic cells thanC. fasciculataparasites. Larvae were used in this experiment since the hemocyte number is significantly higher in larvae than in adults. As shown in Fig. 5A, 1 h after injection, severalC. bombiparasites were internalized by the blood cells, unambiguously identified by the presence of the mitochondrial DNA or kinetoplast, a structure characteristic of this group of flagellate parasites, and by the presence of the flagellum.

Fig. 4. Antimicrobial peptide gene expression following experi-mental intrathoracic parasite-injection. Total RNAs were extracted from control (c) or injected adult ILL 97 flies 24 or 48 h post injection (p.i.) and analyzed by Northern blot. Hybridization signal intensities obtained with diptericin, drosocin and drosomycin cDNA probes were quantified using a Bio-imaging analyzer, and normalized with the sig-nal obtained with the rp49 probe. bact: mixture of Gram-positive and Gram-negative bacteria, C.b:C. bombi, C.f:C. fasciculata. Note that the scale is different in the upper and lower panels.

Fig. 5. Ultrathin section throughDrosophilahemocytes infected with

C. bombi(A) andC. fasciculata(B). Parasites (4.6 nl) were injected into larvae (stage 3). One hour (C.b) or 3 h (C.f) after injection, hemo-lymph was collected and centrifuged to pellet hemocytes and parasites. P: parasite, N: nucleus. Arrows indicate kinetoplast ( ) and flagel-lum (→).

The efficient recognition and phagocytosis of C. bombi

suggested that it might account for the resistance of Dro-sophilato this strain of parasite. We performed a similar experiment with the lethal C. fasciculata species. As shown in Fig. 5B, phagocytosis of C. fasciculata by blood cells was also observed in infected larvae. There-fore, the difference in pathogenicity betweenC.

fascicul-ata and C. bombi cannot be explained by differencial

recognition and phagocytosis by blood cells.

4. Discussion

In this paper, we describe for the first time a model based on the flagellate protozoanCrithidiato study host defense against parasite infection in D. melanogaster. Using antimicrobial peptides as markers to monitor immune response, we present evidence that Crithidia is recognized byDrosophilaupon either systemic infection or local digestive tract infection. Induction is stronger for drosocin (at least ten fold) and it will be interesting to test if this antimicrobial peptide also has antiparasite

(7)

activity, as has been reported for other insect antimicro-bial peptides (Lowenberger et al., 1996; Shahabuddin et al., 1998). In addition, we have shown that depending on the route of infection (e.g. intrathoracic injection vs

per os) and the nature of the infecting microorganism

(e.g. bacteria or Crithidia), different molecules were induced in the hemolymph of infected flies. The differ-ential induction of these peptides by bacterial or Crithi-dia infection strongly argues that these molecules are involved in insect immunity and are not simply stress-related factors.

Because several parasites, including flagellates, carry out part of their development in the insect hemolymph, we have analyzed the immune response of Drosophila

after a systemic infection with Crithidia. Surprisingly, one of theCrithidiaspecies assayed,C. fasciculata, was found to kill Drosophila flies within 5 days after injec-tion in the hemolymph. Similarly, C. fasciculata was lethal when injected into the hemolymph of Glossina

spp., the vector of trypanosomiasis (Ibrahim and Moly-neux, 1987). By contrast,C. bombidid not seem to affect

fascicul-ataandC. bombifor the induction of antimicrobial

pep-tides strongly argues that these molecules are not involved in the control ofCrithidiainfection in the hem-olymph. Humoral immunity also relies on proteolytic cascades leading to melanization (Hoffmann and Reichhart, 1997). In this study, melanization of injected parasites was not observed. Lectins are also involved in insect immunity and they have been shown to play a major role against flagellates (Pereira et al., 1981; Wel-burn et al., 1994; Pimenta et al., 1992; Mello et al., 1999). As with LeishmaniaandTrypanosoma parasites, the membrane of Crithidia spp. is covered uniformly with carbohydrates (Schneider et al., 1996). Hemolymph lectins could interact with parasite carbohydrates and play a role as opsonins to facilitate phagocytosis. Phago-cytosis was observed in our experiments with both

Cri-thidiaspecies, confirming previous results obtained upon

injection of C. fasciculata into Drosophila virilis

(Schmittner and McGhee, 1970). However, we did not observe significant differences in the phagocytosis ofC.

fasciculataandC. bombiwhich could explain the

differ-ence in the outcome of systemic infection by the two species. One possibility is that hemolymph lectins may interact differently with the two parasites due to surface carbohydrate variability. Indeed, lectins can also act as cytotoxic agents as shown inRhodnius prolixusinfected

with Trypanosoma cruzi (Mello et al., 1996).

Using RP-HPLC, we have shown that upon per os

infection antimicrobial peptides were induced in the hemolymph and not locally in the digestive tract. This is somehow surprising since it is well known that in

In Drosophila, the study of transgenic flies expressing

the Green Fluorescent Protein under the control of sev-eral antimicrobial peptide promoters revealed local immune responses in several surface epithelia including the digestive tract (Ferrandon et al., 1998; Ohresser and Imler, unpublished data). Similarly, defensin expression is locally induced in the anterior midgut of Anopheles

gambiae after infection with Plasmodium berghei

(Richman et al., 1997) although Plasmodium gallina-ceuminfection inAedes aegyptidid not induce defensin in the digestive tract of the mosquito (Lowenberger et al., 1999). The fact that we did not observe induction of known antimicrobial peptides in the digestive tract in our RP-HPLC and RT-PCR experiments does not exclude that other, yet-to-be-discovered molecules are induced in the gut following Crithidia infection. This raises the question of the mechanism by which digestive tract infection triggers antimicrobial peptide expression by the fat body. One possibility is that heavy infection occasionally results in tissue damage or migration of parasites into the hemolymph (Schaub, 1994). This hypothesis was investigated by collecting hemolymph from flies infected per os with Crithidia spp. After microscopic observation, noCrithidiaspp. was detected in the hemolymph 24 h after injection (data not shown). Another possible mechanism could involve cytokine-like molecules. In this model, interaction of flagellate para-sites with the digestive tract epithelium, possibly through the formation of hemidesmosomes (Ismaeel, 1994), would induce the synthesis of such cytokine-like mol-ecules, that would then activate cell surface receptors on fat body cells and induce antimicrobial peptide synthesis. Interestingly, we report here the identification of three molecules of molecular weight 2898 Da, 4346 Da and 7192 Da which are induced in the hemolymph afterper os infection with bacteria or Crithidia, but not after injection of bacteria. These molecules represent good candidates for such signaling messengers between the gut and the fat body and their biochemical characteriz-ation is under progress.

In conclusion, this paper describes the immune response of Drosophila to Crithidia infection and pro-vides a new model to better understand the molecular basis of host defense against flagellate parasites in insects using the powerful genetics of Drosophila. Our results support previous work showing that insects can discriminate between pathogens during an infection (Lemaitre et al., 1997). We are in parallel extending this study by working with two other flagellates,

(8)

Trypano-soma spp. and Leishmania spp., which are transmitted to humans and animals by Glossina spp. and

Phle-botomus spp., respectively. Altogether, these

experi-mental models should allow us to identify and charac-terize novel molecules which may be involved in the control of parasitic infections in Diptera. In this regard, the identification of a molecule of 3175 Da which is exclusively induced in the hemolymph of Drosophila

after a per os infection with Crithidia is particularly encouraging.

Acknowledgements

We would like to thank Prof. J.A. Hoffmann and Prof. B. Pesson for support and interest and Dr R. Lanot for the preparation of larvae for electron microscopy. This work was supported by grants from the Training and Mobility of Researchers (TMR) Programme of the Euro-pean Union and by the CNRS, Centre National de la Recherche Scientifique, France.

References

Beerntsen, B., Severson, D.W., Christensen, B.M., 1994. Aedes aegypti: characterization of a hemolymph polypeptide expressed during melanotic encapsulation of filarial worms. Exp. Parasitol. 79, 312–321.

Boman, H.G., 1998. Gene-encoded peptide antibiotics and the concept of innate immunity: an update review. Scand. J. Immunol. 48, 15–25.

Brey, P.T., Lee, W.J., Yamakawa, M., Koizumi, Y., Perrot, S., Fran-c¸ois, M., Ashida, M., 1993. Role of the integument in insect immunity: epicuticular abrasion and induction of cecropin synthesis in cuticular epithelial cells. Proc. Natl. Acad. Sci. USA 90, 6275–6279.

Bulet, P., 1999. Les peptides antimicrobiens de la drosophile. Me´decine/sciences 15, 23–29.

Bulet, P., Uttenweiler-Joseph, S., 1999. A MALDI-TOF mass spec-trometry approach to investigate the defense reactions in Droso-phila melanogaster, an insect model for the study of innate immun-ity. In: Kamp, R.M., Kyriakidis, D., Choli-Papadopoulou, Th. (Eds.), Proteome and Protein Analysis. Springer-Verlag, Berlin. Dimopoulos, G., Seeley, D., Wolf, A., Kafatos, F.C., 1998. Malaria

infection of the mosquitoAnopheles gambiae activates immune-responsive genes during critical transition stages of the parasite life cycle. EMBO J. 17, 6115–6123.

Ferrandon, D., Jung, A., Criqui, M.C., Lemaitre, B., Uttenweiler-Joseph, S., Michaut, L., Reichhart, J.M., Hoffmann, J.A., 1998. A drosomycin-GFP reporter transgene reveals a local immune response inDrosophilathat is not dependent on the Toll pathway. EMBO J. 17, 1217–1227.

Ganz, T., Lehrer, R., 1998. Antimicrobial peptides of vertebrates. Curr. Opin. Immunol. 10, 41–44.

Hoffmann, J.A., Reichhart, J.M., 1997.Drosophilaimmunity. Trends Cell Biol. 7, 309–316.

Hoffmann, J.A., Kafatos, F.C., Janeway, C.A., Ezekowitz, R.A.B., 1999. Phylogenetic perspectives in innate immunity. Science 284, 1313–1318.

Ibrahim, E.A., Molyneux, D.H., 1987. Pathogenicity ofCrithidia fas-ciculatain the haemocele ofGlossina. Acta Trop. 44, 13–22.

Ismaeel, A.Y., 1994. Studies on host–parasite relationships of trypano-somatid flagellates inDrosophila, and Leishmaniain Phlebotom-ines sandflies. Ph.D. Thesis, University of Liverpool, UK. Kaaya, G.P., Otieno, L.H., Darji, N., Alemu, P., 1986. Defense

reac-tions ofGlossina morsitans morsitansagainst different species of bacteria andTrypanosoma brucei brucei. Acta Trop. 43, 31–42. Lackie, A.M., 1988. Immune mechanisms in insects. Parasitol. Today

4, 98–105.

Lehane, M.J., Wu, D., Lehane, S., 1997. Midgut specific immune mol-ecules are produced by the blood sucking insectStomoxys calci-trans. Proc. Natl. Acad. Sci. USA 94, 11502–11507.

Lemaitre, B., Nicolas, E., Michaut, L., Reichhart, J.-M., Hoffmann, J.A., 1996. The dorsoventral regulatory gene cassette spaetzle/Toll/cactus controls the potent antifungal response in Dro-sophilaadults. Cell 86, 973–983.

Lemaitre, B., Reichhart, J.-M., Hoffmann, J.A., 1997.Drosophilahost defense: differential induction of antimicrobial peptide genes after infection by various classes of microorganisms. Proc. Natl. Acad. Sci. USA 94, 14614–14619.

Lowenberger, C.A., Ferdig, M.T., Bulet, P., Khalili, S., Hoffmann, J.A., Christensen, B.M., 1996.Aedes aegypti: induced antibacterial proteins reduce establishment and development ofBrugia malayi. Exp. Parasitol. 83, 191–201.

Lowenberger, C.A., Kamal, S., Chiles, J., Paskewitz, S., Bulet, P., Hoffmann, J.A., Christensen, B.M., 1999. Mosquito–Plasmodium

interactions in response to immune activation of the vector. Exp. Parasitol. 91, 59–69.

Mello, C.B., Azambuja, P., Garcia, E.S., Ratcliffe, N.A., 1996. Differ-ential in vitro and in vivo behavior of three strains ofTrypanosoma cruziin the gut and hemolymph ofRhodnius prolixus. Exp. Parasi-tol. 82, 112–121.

Mello, C.B., Nigam, Y., Garcia, E.S., Azambuja, P., Newton, R.P., Ratcliffe, N.A., 1999. Studies on a haemolymph lectin isolated fromRhodnius prolixusand its interaction withTrypanosoma rang-eli. Exp. Parasitol. 91, 289–296.

Molyneux, D.H., Killick-Kendrick, R., 1987. Morphology, ultrastruc-ture and life cycles in The Leishmaniases. In: Peters, W., Killick-Kendrick, R. (Eds.), Biology and Medicine, vol. I. Academic Press, London, pp. 140–161.

Pereira, M.A.E., Andrade, A., Ribeiro, J.M., 1981. Lectins of distinct specificity inRhodnius prolixusinteract selectively with Trypano-zoma cruzi. Science 211, 597–600.

Pimenta, F.P., Turco, S.J., McConville, M.J., Lawyer, P.G., Perkins, P.V., Sacks, D., 1992. Stage-specific adhesion ofLeishmania pro-mastigotes to the sandfly midgut. Science 256, 1812–1815. Richman, A.M., Kafatos, F.C., 1995. Immunity to eukaryotic parasites

in vector insects. Curr. Opin. Immunol. 8, 14–19.

Richman, A.M., Dimopoulos, G., Seeley, D., Kafatos, F.C., 1997.

Plasmodium activates the innate immune response ofAnopheles gambiaemosquitoes. EMBO J. 16, 6114–6119.

Russell, V., Dunn, P.E., 1996. Antibacterial proteins in the midgut of Manduca sexta during metamorphosis. J. Insect. Physiol. 42, 65–71.

Schaub, G.A., 1994. Pathogenicity of trypanosomatids on insects. Para-sitol. Today 10, 463–468.

Shahabuddin, M., Fields, I., Bulet, P., Hoffmann, J.A., Miller, L.H., 1998.Plasmodium gallinaceum: differential killing of some mos-quito stages of the parasite by insect defensin. Exp. Parasitol. 89, 103–112.

Schmittner, S.M., McGhee, R.B., 1970. Host specificity of various species ofCrithidiaLe´ger. J. Parasitol. 56, 684–693.

Schneider, P., Treumann, A., Milne, K.G., McConville, M.J., Zitz-mann, N., Ferguson, M.A., 1996. Structural studies on a lipoarabin-ogalactan ofCrithidia fasciculata. Biochem. J. 313, 963–971. Uttenweiler-Joseph, S., Moniatte, M., Lagueux, M., van Dorsselaer,

A., Hoffmann, J.A., Bulet, P., 1998. Differential display of peptides induced during the immune response of Drosophila: a

(9)

matrix-assisted laser desorption ionization time-of-flight mass spec-trometry study. Proc. Natl. Acad. Sci. USA 95, 11342–11347. Wallace, F.G., 1966. The trypanosomatid parasites of insects and

arachnids. Exp. Parasitol. 18, 124–193.

Wallace, F.G., 1979. Biology of theKinetoplastidaof arthopods. In:

Lumsden, W.H.R., Evans, D.A. (Eds.), Biology ofKinetoplastida, vol. 2. Academic Press, London, pp. 213–235.

Welburn, S.C., Maudlin, I., Molyneux, D.H., 1994. Midgut lectin activity and sugar specificity in teneral and fed tse-tse. Med. Vet. Entomol. 8, 81–87.

(1)

3.2. Per osinfection with flagellates triggers a systemic immune response

Surface epithelia have been shown to synthesize anti-microbial peptides both in mammals (reviewed in Ganz and Lehrer, 1998) and in various insect species, namely

Bombyx mori (Brey et al., 1993), Manduca sexta

Crithidia induces antimicrobial peptide synthesis in the digestive tract (local response) or in the fat body (systemic response), we performed a separate analysis by reverse phase HPLC (RP-HPLC) and mass spec-trometry of acidic extracts of gut tissue and hemolymph from individual flies afterper osinfection withCrithidia

spp. or with bacteria. As shown in Fig. 2, infection by both C. fasciculata and bacteria induced a systemic immune response illustrated by the appearance of droso-cin and drosomydroso-cin (peaks 1 and 2, respectively) in the hemolymph of infected flies, 24 h after infection. The

The per os infection with Crithidia spp. or bacteria also induced molecules distinct from the known anti-microbial peptides (Fig. 2A). Mass spectrometry analy-sis of peak B yielded several masses ranging from approximately 3 kDa to 7 kDa, one of which (mol. wt 3175 Da) was specific for Crithidia infection (Table 1). The corresponding molecules are still under investi-gation. Interestingly, peak B was observed only in per os infected flies, and was not detected in flies injected with bacteria. Peaks A and C of Fig. 2A were only induced by bacteria, both afterper osadministration and injection. These data demonstrate that not only the nat-ure of infecting microorganisms but also the route of infection induce different molecules in the hemolymph

(2)

Table 1

Specificity of the immune response inDrosophilainjected into the thorax with bacteria or infectedper oswith different pathogens (bacteria and

Crithidiaspp.)a

Per os C. fasciculata

Peak Bacteria injection Per osbacteria infection Per os C. bombiinfection

infection

A DIM 1b _{DIM 1}

2 2

DIM 4 DIM 4

B 2 2899.0 2899.4 2899.0

3176.6 3175.2

4347.3 4347.8 4347.3

7193.0 7193.5 7193.0

C DIM 13 DIM 13 2 2

a_{The RP-HPLC fractions corresponding to induced peaks A, B and C were analyzed by MALDI-TOF-MS. DIMs 1 (1666 Da), 4 (1722} Da), 13 (2651 Da) were previously described (Uttenweiler-Joseph et al., 1998). Measured molecular masses (MH+) are indicated for unknown induced molecules.

b _{DIM —}_Drosophila_{Immune-induced Molecules.}

of Drosophila. Using a similar methodology, we have analyzed gut tissue from bacteria- andCrithidia-infected field strain Drosophila. We were unable to detect the presence of antimicrobial peptides in this type of extract 24 h after aper osinfection. These data were confirmed by reverse transcription-PCR (RT-PCR) analysis which showed expression of drosocin and diptericin in dis-sected fat bodies from per os infected flies, but not in dissected guts (data not shown).

3.3. Injection of C. fasciculata kills flies

Some flagellate parasites such as certainTrypanosoma

spp. (Schaub, 1994). This has prompted us to analyze the response of flies to injection of parasites. First, we recorded the survival of field strain flies to intrathoracic injection of C. fasciculata and C. bombi over 6 days (Fig. 3). In contrast to per os administration, which did not induce any lethality, injection ofC. fasciculatakilled all flies within 4 days post-injection. Injection of C. bombi was not harmful and up to 80% of injected flies had survived 6 days after injection of the parasite or PBS. Interestingly, dissection ofC. bombi-infected flies showed few viable (e.g. motile) parasites 4–6 days after injection while C. fasciculata-infected flies contained large numbers of highly motile parasites.

3.4. Injection of Crithidiaparasites induces a weak systemic antimicrobial response

C. fasciculata or C. bombi results from a difference in the level of induction of antimicrobial peptides. To investigate this, we analyzed the expression of the corre-sponding genes in flies 1 and 2 days after injection of both parasites, and compared this response with that

(3)

3.5. Phagocytosis of Crithidiaparasites by hemocytes

Fig. 5. Ultrathin section throughDrosophilahemocytes infected with

The efficient recognition and phagocytosis of C. bombi

4. Discussion

(4)

per os) and the nature of the infecting microorganism (e.g. bacteria or Crithidia), different molecules were induced in the hemolymph of infected flies. The differ-ential induction of these peptides by bacterial or Crithi-dia infection strongly argues that these molecules are involved in insect immunity and are not simply stress-related factors.

Because several parasites, including flagellates, carry out part of their development in the insect hemolymph, we have analyzed the immune response of Drosophila

spp., the vector of trypanosomiasis (Ibrahim and Moly-neux, 1987). By contrast,C. bombidid not seem to affect

Drosophila, thus suggesting a differential recognition of the two parasite species. The reason for this difference is still not clear at present. The fact that we could not observe any significant differences between C. fascicul-ataandC. bombifor the induction of antimicrobial pep-tides strongly argues that these molecules are not involved in the control ofCrithidiainfection in the hem-olymph. Humoral immunity also relies on proteolytic cascades leading to melanization (Hoffmann and Reichhart, 1997). In this study, melanization of injected parasites was not observed. Lectins are also involved in insect immunity and they have been shown to play a major role against flagellates (Pereira et al., 1981; Wel-burn et al., 1994; Pimenta et al., 1992; Mello et al., 1999). As with LeishmaniaandTrypanosoma parasites, the membrane of Crithidia spp. is covered uniformly with carbohydrates (Schneider et al., 1996). Hemolymph lectins could interact with parasite carbohydrates and play a role as opsonins to facilitate phagocytosis. Phago-cytosis was observed in our experiments with both Cri-thidiaspecies, confirming previous results obtained upon injection of C. fasciculata into Drosophila virilis

(Schmittner and McGhee, 1970). However, we did not observe significant differences in the phagocytosis ofC. fasciculataandC. bombiwhich could explain the differ-ence in the outcome of systemic infection by the two species. One possibility is that hemolymph lectins may interact differently with the two parasites due to surface carbohydrate variability. Indeed, lectins can also act as cytotoxic agents as shown inRhodnius prolixusinfected with Trypanosoma cruzi (Mello et al., 1996).

Using RP-HPLC, we have shown that upon per os

infection antimicrobial peptides were induced in the hemolymph and not locally in the digestive tract. This is somehow surprising since it is well known that in

ver-tebrates, epithelia participate in the host defense by secreting locally active antimicrobial peptides (Ganz and Lehrer, 1998). Similarly, in invertebrates, a local consti-tutive expression of defensin has been demonstrated in the digestive tract of Stomoxys spp., and expression of this antibacterial peptide increased following a lipopoly-saccharide-containing blood meal (Lehane et al., 1997). In Drosophila, the study of transgenic flies expressing the Green Fluorescent Protein under the control of sev-eral antimicrobial peptide promoters revealed local immune responses in several surface epithelia including the digestive tract (Ferrandon et al., 1998; Ohresser and Imler, unpublished data). Similarly, defensin expression is locally induced in the anterior midgut of Anopheles gambiae after infection with Plasmodium berghei

(5)

Trypano-soma spp. and Leishmania spp., which are transmitted to humans and animals by Glossina spp. and Phle-botomus spp., respectively. Altogether, these experi-mental models should allow us to identify and charac-terize novel molecules which may be involved in the control of parasitic infections in Diptera. In this regard, the identification of a molecule of 3175 Da which is exclusively induced in the hemolymph of Drosophila